From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
---|---|
To: | "Hitoshi Harada" <umi(dot)tanuki(at)gmail(dot)com> |
Cc: | "Bruce Momjian" <bruce(at)momjian(dot)us>, PostgreSQL-development <pgsql-hackers(at)postgresql(dot)org> |
Subject: | Re: tuplestore potential performance problem |
Date: | 2009-01-15 14:51:00 |
Message-ID: | 5413.1232031060@sss.pgh.pa.us |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
"Hitoshi Harada" <umi(dot)tanuki(at)gmail(dot)com> writes:
> 2009/1/15 Bruce Momjian <bruce(at)momjian(dot)us>:
>> Has this been addressed?
> It is mentioned at
> http://archives.postgresql.org/pgsql-hackers/2008-12/msg01849.php
> but not solved yet. It seems to me that to solve this the tuplestore's
> inside design should be changed much. In-file state doesn't use memory
> any more but it should be re-used for writing buffer, whereas the
> current desgin uses BufFile to do it, which causes tell/seek overhead
> for repeated put/get operation. And this is not for 8.4, I guess.
I was thinking of something a lot simpler, like just allowing BufFile
to maintain separate buffers for read and write pointers. But in any
case it's not likely to get done for 8.4.
regards, tom lane
From | Date | Subject | |
---|---|---|---|
Next Message | Heikki Linnakangas | 2009-01-15 14:53:47 | Re: cleanup smgr.c of tablespace call |
Previous Message | Heikki Linnakangas | 2009-01-15 14:25:29 | Re: BUG #4566: pg_stop_backup() reports incorrect STOP WAL LOCATION |